Content uploaded by Raymond B Huey
Author content
All content in this area was uploaded by Raymond B Huey on Jun 03, 2014
Content may be subject to copyright.
LETTER doi:10.1038/nature09407
Global metabolic impacts of recent climate warming
Michael E. Dillon
1
, George Wang
2
{& Raymond B. Huey
2
Documented shifts in geographical ranges
1,2
, seasonal phenology
3,4
,
community interactions
5
, genetics
3,6
and extinctions
7
have been
attributed to recent global warming
8–10
. Many such biotic shifts
have been detected at mid- to high latitudes in the Northern
Hemisphere
4,9,10
—a latitudinal pattern that is expected
4,8,10,11
because warming is fastest in these regions
8
. In contrast, shifts in
tropical regions are expected tobe less marked
4,8,10,11
because warm-
ing is less pronounced there
8
. However, biotic impacts of warming
are mediated through physiology, and metabolic rate, which is a
fundamental measure of physiological activity and ecological
impact, increases exponentially rather than linearly with tem-
perature in ectotherms
12
. Therefore, tropical ectotherms (with
warm baseline temperatures) should experience larger absolute
shifts in metabolic rate than the magnitude of tropical temperature
change itself would suggest, but the impact of climate warming on
metabolic rate has never been quantified on a global scale. Here we
show that estimated changes in terrestrial metabolic rates in the
tropics are large, are equivalent in magnitude to those in the north
temperate-zone regions, and are in fact far greater than those in
the Arctic, even though tropical temperature change has been rela-
tively small. Because of temperature’s nonlinear effects on meta-
bolism, tropical organisms, which constitute much of Earth’s
biodiversity, should be profoundly affected by recent and projected
climate warming
2,13,14
.
Global warming is probably having profound and diverse effects on
organisms
1
–
11
. Organisms living at mid- to high latitudes in the
Northern Hemisphere are predicted to be the most affected by climate
warming
4,8,10,11
, because temperatures have risen most rapidly there
8
.
Indeed, the vast majority of biotic impacts of warming have been
documented in this region, but few studies have yet searched for
impacts in other areas, especially the tropics
2,4,8,10,13,14
. One way to
circumvent this geographical sampling bias is to use temperature data
with broad geographical coverage to predict global patterns of physio-
logical responses to observed temperature change
13
. Metabolic rate is a
heuristic metric here because it is a fundamental physiological index of
an organism’s energetic and material needs, its processing capacity and
its ecological impact
12
.
Metabolic rates of ectotherms depend principally on body mass (m)
and body temperature (T), as described by a fundamental equation
12
:
B(m,T) 5b
0
m
3/4
e
2E/kT
(1)
where Bis metabolic rate, b
0
is an empirically derived and taxon-
specific normalization constant, mis body mass, Eis the average activation
energy for biochemical reactions of metabolism, Tis body temperature (in
Kelvin), and kis the Boltzmann constant. When standardized for mass,
this equation enables metabolic comparisons among different sized
organisms
12
. These mass-normalized metabolic rates are proportional
to the ‘Boltzmann factor’ (e
2E/kT
;thefamiliar‘Q
10
’ effect in physiology
is an approximation of the Boltzmann factor)
15
.
Although the thermodynamic and statistical validity of equation (1)
is debated
16–18
, it provides a useful approximationof metabolic rates
12,15
for exploratory macrophysiological investigations
16,19
. In the context of
climate warming, it predicts that metabolism will shift more in res-
ponse to a unit change in temperature at high temperature than at low
temperature
15
, at least over biologically common and non-stressful
temperatures (,0uCto,40 uC; see Methods)
12
.
To estimate geographical patterns of warming-induced changes in
metabolic rates of terrestrial ectotherms, we compiled high-frequency
temperature data for the period of 1961 to 2009 for 3,186 weather
stations across the world (,500 million temperature measurements;
Methods and Supplementary Fig. 1)
20
. We derived average values of
E(0.69) and of b
0
(23.66) from empirical estimates for diverse
ectotherms (Supplementary Table 1)
12
. We substituted these ‘average
ectotherm’ values into equation (1) to estimate mass-normalized meta-
bolic rates (Bm
23/4
) from global temperature data. Because metabolic
rate varies nonlinearly with temperature, calculating mean metabolic
rates from mean temperatures is inappropriate (the ‘fallacy of the
averages’; see Methods and Supplementary Fig. 2)
15
. Therefore we
estimated metabolic rate for each temperature measurement and sub-
sequently determined average temperature and average metabolic rate
for each station during the Intergovernmental Panel on Climate
Change (IPCC) standard reference period (1961–1990) and for all
five-year intervals from 1980 to 2009. To account for non-uniform
distribution of stations and to enable comparisons among latitudinal
regions, we determined averages for all stations within 5ulatitude by 5u
longitude grid cells and then area-corrected grid-cell means and standard
errors of temperature measurements and metabolic rate estimates for
each region (Methods).
Temperature changes since 1980 in this data set are consistent with
recent findings
8
: temperatures rose fastest in the Arctic, somewhat less
quicklyin the north temperatezones, and more slowlyin the tropics, but
remained essentially unchanged in the south temperate zone (Fig. 1a).
Predicted absolute changes in metabolic rates show a markedly
different pattern: metabolic rates increased most quickly in the tropics
and north temperate zones, and less so in the Arctic (Fig. 1b). In fact,
the latitudinal ordering of changes in temperature since 1980 fails to
predict the latitudinal ordering of changes in metabolic rate (P50.68),
even when a powerful ordered-heterogeneity test is used
21
. The pre-
dicted increase in metabolism in the tropics was large, despite the small
rise in temperature there (Fig. 1a), because tropical warmingtook place
in an environment that was initially warm.
Absolute changes in metabolic rates determine an organism’s total
energy use and thus the impacts of climate change on ecosystem-level
processes, but per cent changes in metabolic rates are nonetheless
relevant to the impacts of climate change on individual organisms
22
.
Such relative changes in metabolic rates (expressed as per cent of the
standard reference period on a per-station basis) closely match tem-
perature changes (Fig. 1c), indicating that impacts on individual
ectotherms have probably been relatively large in the Arctic and north
temperate zones.
To evaluate whether the patterns described earlier (Fig. 1b) are robust
to our use of average values of Eand b
0
, we re-ran analyses using esti-
mates of Eand b
0
specific to diverse ectotherm taxa (Supplementary
Table 1)
12
. Large effects of recent climate warming on metabolic rates
are predicted for invertebrates, amphibians and reptiles in equatorial
1
Department of Zoology and Physiology, University of Wyoming, Laramie, Wyoming 82071, USA.
2
Department of Biology, Box 351800, University of Washington, Seattle, Washington 98195, USA. {Present
address: Max Planck Institute for Developmental Biology, Tu
¨bingen 72076, Germany.
704 | NATURE | VOL 467 | 7 OCTOBER 2010
Macmillan Publishers Limited. All rights reserved
©2010
West Africa, the Caribbean and Central America, Ecuador, eastern
equatorial Brazil, and the Persian Gulf region (Fig. 2c–e). However, we
emphasize that weather station coverage in some of these regions is
sparse and each taxonomic group is not found in all geogr aphical regions.
Overall, general patterns in Fig. 1bare robust for different taxa (Fig. 2b–e):
the largest predicted absolute shifts in metabolic rate for all taxa are in
the tropics. Nevertheless, small differences in the relationship between
metabolism and temperature (that is, Eand b
0
) can alter the magnitude
of the effects of climate warming on organism physiology (Fig. 2).
The patterns under discussion are for mass-normalized metabolic
rates, but the magnitude of metabolic shift will necessarily differ for small
versus large ectotherms. Of course, absolute shifts will be greater for larger
ectotherms, but equation (1) indicates that mass-specific metabolic rates
of small ectotherms will show larger increases (Supplementary Fig. 3).
Several assumptions underlie the patterns shown in Fig. 1b, c. The
exponent (3/4) for metabolic rate as a function of mass is debated
16–18
,
but reasonable shifts of this exponent for given taxa will only alter the
heights of all latitudinal lines, not their relative ordering. We assume
that surface air temperatures approximate ectotherm body tempera-
tures; therefore, our metabolic estimates apply to thermoconforming
and exposed ectotherms. This is reasonable for small ectotherms living
in shaded environments
23
, but less so for large ectotherms that live in
thermally heterogeneous environments, where behavioural thermo-
regulation is possible, or for organisms that spend extensive periods
in retreats
24
. Also, we assume that the coefficients (E,b
0
) of equation
(1) are independentof latitude.However, somehigh-latitude ectotherms
have relatively elevated metabolic rates; and this is thought to represent
an evolutionary metabolic compensation for the physiologically depres-
sing effects of low body temperature
16,25,26
. With reference to equation
(1), metabolic compensation would be indicated
26
by latitudinal
increases in b
0
and/or in E. In fact, the patterns of metabolic responses
shown in Fig. 1b hold even when we shift these parameters over a large
−0.5
0
0.5
1.0
1.5
2.0
1980 1990 2000 2010
a
Change in temperature (°C)
Tropical
North temperate
South temperate
Arctic
Tropical
North temperate
South temperate
Arctic
Tropical
North temperate
South temperate
Arctic
−0.01
0
0.01
0.02
0.03
0.04
0.05
1980 1990 2000 2010
b
Change in metabolic rate (mW g–3/4)
YearYear Year
−5
0
5
10
15
20
25
1980 1990 2000 2010
c
Change in metabolic rate (%)
Figure 1
|
Global changes in temperature and in metabolic rates since 1980.
a, Changes in mean temperature (5-year averages) for Arctic (n5100 grid
cells), north temperate (n5356), south temperate (n551) and tropical
(n5169) regions. b, Predicted absolute changes in mass-normalized metabolic
rates by geographical region. c, Predicted relative changes in mass-normalized
metabolic rates. Both temperature and metabolic rate are expressed as
differences from the standard reference period (1961–1990), calculated on a
per-station basis, on the basis of Eand b
0
for an average ectotherm
(Supplementary Table 1). Data points are means 6s.e.m. of area-corrected,
gridded weather-station data (Methods).
60° S
30° S
0°
30° N
60° N
60° N
60° N 60° N
60° N
180° W
120° W
60° W
0°
60° E
120° E
180° E
60° S
30° S
0°
30° N
180° W
120° W
60° W
0°
60° E
120° E
180° E
60° S
30° S
0°
30° N
180° W
120° W
60° W
0°
60° E
120° E
180° E
60° S
30° S
0°
30° N
180° W
120° W
60° W
0°
60° E
120° E
180° E
60° S
30° S
0°
30° N
180° W
120° W
60° W
0°
60° E
120° E
180° E
aChange in temperature
−1.4 °C
3.4 °C
b
Change in metabolic rate
Unicellular organisms
−0.08
mW g−3/4
0.15
mW g−3/4
cInvertebrates
dAmphibians eReptiles
Figure 2
|
Predicted changes in metabolic rates
of diverse terrestrial ectotherms. a, Difference in
temperature between 1961–1990 and 2005–2009,
with scale bar shown on right. b–e, Difference in
mass-normalized metabolic rates (predicted) for
the same period for four terrestrial ectothermic
animal taxa for which empirical estimates of Eand
b
0
are available (Supplementary Table 1)
12
. Colour
bar to right of bindicates scale for b–e. Grey
shading indicates grid cells with no temperature
data.
LETTER RESEARCH
7 OCTOBER 2010 | VOL 467 | NATURE | 705
Macmillan Publishers Limited. All rights reserved
©2010
range of biologically reasonable values
26
to simulate extreme metabolic
compensation at high latitude (Methods and Supplementary Fig. 4).
Our analyses indicate that warming during the past three decades has
had its biggest absolute impacts on metabolic rates in tropical and north
temperate zones (Fig. 1b). The outlook for future warming is less clear.
Without predictions of future daily and seasonal temperature cycles (not
merely of mean annual temperatures), we cannot directly estimate
future metabolic changes without violating the fallacy of averages
15
.
Nevertheless, our analyses of recent temperature data indicate that even
when the temperature shifts in the north temperate region are more than
double those in the tropics (Fig. 1a), absolute shifts in metabolic rates are
similar in the two regions (Fig. 1b). If this pattern holds, projected
increases in median surface air temperature by the end of the twenty-
first century for the two regions (3.5–4.0 uC in the tropics,and 4.0–5.5 uC
in the north temperate zone)
8
should cause roughly similar absolute
increases in metabolic rates of tropical and north temperate organisms.
Recent studies using diverse physiological and biophysical ap-
proaches indicate that tropical ectotherms may be particularly vulner-
able to climate warming
2,7,13,14,24,27
, even though observed and predicted
tropical warming is relatively small
8
. Our estimates suggest that tropical
ectotherms are also experiencing large increases in metabolic rate (Fig.
1b). Such increases will have physiological and ecological impacts:
warmed tropical ectotherms will have an increased need for food and
increased vulnerability to starvation unless food resources increase,
possible reduced discretionary energy for reproduction
22
, increased
rates of evaporative water loss in dry environments and altered demo-
graphies
13
. Larger increases in metabolic rates of tropical soil biota
may explain larger absolute changes in tropical soil respiration
28
.
Furthermore, metabolic increases should alter food web dynamics,
leading to elevated rates of herbivory and predation, as well as changes
in the spread of insect-borne tropical diseases
29
. Because the tropics are
the centre of Earth’s biodiversity and its chief engine of primary pro-
ductivity, the relatively large effects of temperature change on the meta-
bolism of tropical ectotherms may have profound local and global
consequences.
METHODS SUMMARY
We obtained hourly temperature records from 22,486 weather stations spread
across the world
20
, but then included only stations that sampled throughout the
IPCC standard reference period (1961–1990)
8
as well as 1991–2009, in all seasons
and on average at leastevery six hours.We also excludedfive Antarcticstations,such
that 3,186 stations remained. Geographical coverage is uneven (Supplementary
Fig. 1), but all regions are well represented in this restricted data set (Supplemen-
tary Table 2). Furthermore, including data from 5,561 stations with data from 1961
to 2009 (but with no other limitations; Supplementary Table 2), does not alter our
conclusions (Supplementary Fig.5). To correct for the unevenspatial distribution of
stations (Supplementary Figs 1 and 5A), we computed mean temperatures and
metabolicrates for all stations within 5ulatitude by 5ulongitude cells. We calculated
means and standard errors for latitudinal regions by weighting grid-cell means by
the interpolated rectangular mid-cell areas
30
. To estimate whether metabolic com-
pensation at high latitude might alter patterns,we recalculated metabolic rates after
substituting extreme values of Eand of b
0
. Specifically, we used veryhigh values of E
(0.76)and b
0
(26.85)for the north temperate and Arctic areas,but very low valuesof
E(0.50) andb
0
(15.68)for tropical areas(see Supplementary Table 1). Such extreme
metabolic compensation (these values span most of the known range of E)
26
does
not alter our conclusions (Supplementary Fig. 4).
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 5 June; accepted 10 August 2010.
1. Parmesan, C. & Yohe,G. A globally coherent fingerprint of climate changeimpacts
across natural systems. Nature 421, 37–42 (2003).
2. Colwell, R. K., Brehm, G., Cardelus, C. L., Gilman, A. C. & Longino, J. T. Global
warming, elevational range shifts, and lowland biotic attrition in the wet tropics.
Science 322, 258–261 (2008).
3. Bradshaw, W. & Holzapfel, C. Genetic shift in photoperiodic response correlated
with global warming. Proc. Natl Acad. Sci. USA 98, 14509–14511 (2001).
4. Parmesan, C. Influences of species, latitudesand methodologies on estimates of
phenologicalresponsetoglobalwarming.Glob.ChangeBiol. 13, 1860–1872(2007).
5. Both, C., vanAsch, M., Bijlsma, R., van denBurg, A. & Visser, M. Climate changeand
unequal phenological changes across four trophic levels: constraints or
adaptations? J. Anim. Ecol. 78, 73–83 (2009).
6. Umina, P. A., Weeks, A. R., Kearney, M. R., McKechnie, S. W. & Hoffmann, A. A. A
rapid shift in a classic clinal pattern in Drosophila reflecting climate change.
Science 308, 691–693 (2005).
7. Sinervo, B. et al. Erosion of lizard diversity by climate change and altered thermal
niches. Science 328, 894–899 (2010).
8. IPCC. Climate Change 2007: Impacts, Adaptation, and Vulnerability (Cambridge
Univ. Press, 2007).
9. Walther, G.-R. et al. Ecological responses to recent climate change. Nature 416,
389–395 (2002).
10. Rosenzweig, C. et al. Attributing physical and biological impacts to anthropogenic
climate change. Nature 453, 353–357 (2008).
11. Root, T. L. et al. Fingerprints of global warming on wild animals and plants. Nature
421, 57–60 (2003).
12. Gillooly, J. F., Brown, J. H., West, G. B., Savage, V. M. & Charnov, E. L. Effects of size
and temperature on metabolic rate. Science 293, 2248–2251 (2001).
13. Deutsch, C. A. et al. Impacts of climate warming on terrestrial ectotherms across
latitude. Proc. Natl Acad. Sci. USA 105, 6668–6672 (2008).
14. Pounds, J., Fogden, M. & Campbell, J. Biological response to climate change on a
tropical mountain. Nature 398, 611–615 (1999).
15. Savage, V. M. Improved approximations to scaling relationships for species,
populations,and ecosystems across latitudinaland elevational gradients.J. Theor.
Biol. 227, 525–534 (2004).
16. Clarke, A. Temperature and the metabolic theory of ecology. Funct. Ecol. 20,
405–412 (2006).
17. Downs, C. J., Hayes, J. P. & Tracy, C. R. Scaling metabolic rate with body massand
inverse bodytemperature: a test of theArrhenius fractal supplymodel. Funct. Ecol.
22, 239–244 (2008).
18. O’Connor, M. P. et al. Reconsidering themechanistic basis of the metabolic theory
of ecology. Oikos 116, 1058–1072 (2007).
19. Martı
´nez del Rio, C. Metabolic theory or metabolic models? Trends Ecol. Evol. 23,
256–260 (2008).
20. Lott, N., Baldwin, R. & Jones, P. The FCC Integrated Surface Hourly Database, A New
Resourceof Global ClimateData. Æhttp://www1.ncdc.noaa.gov/pub/data/techrpts/
tr200101/tr2001-01.pdfæ(National Climatic Data Center, 2001).
21. Rice, W. R. & Gaines, S. D. Extending non-directionalheterogeneity teststo evaluate
simplyorderedalternativehypotheses.Proc. Natl Aca d.Sc i.US A91 , 225–226(1994).
22. Dunham, A. E. in Biotic Interactions and Global Change (eds Kareiva, P. M.,
Kingsolver, J. G. & Huey, R. B.) 95–119 (Sinauer, 1993).
23. Hertz, P. E., Huey, R. B. & Stevenson, R. D. Evaluating temperature regulation by
field-active ectotherms: the fallacy of the inappropriate question. Am. Nat. 142,
796–818 (1993).
24. Kearney, M., Shine, R. & Porter,W. P. The potential for behavioral thermoregulation
to buffer ‘‘cold-blooded’’ animals against climate warming. Proc. Natl Acad. Sci.
USA 106, 3835–3840 (2009) .
25. Po
¨rtner, H. Physiological basis of temperature-dependent biogeography:
tradeoffs in muscledesign and performance in polar ectotherms. J. Exp. Biol. 205,
2217–2230 (2002).
26. Irlich, U. M., Terblanche, J. S., Blackburn, T. M. & Chown, S. L. Insect rate–
temperature relationships: environmental variation and the metabolic theory of
ecology. Am. Nat. 174, 819–835 (2009).
27. Wake, D. B. & Vredenburg, V. T. Are we in the midst of the sixth mass extinction? A
view from the world of amphibians. Proc. Natl Acad. Sci. USA 105, 11466–11473
(2008).
28. Bond-Lamberty, B. & Thomson,A. Temperature-associated increases in the global
soil respiration record. Nature 464, 579–582 (2010).
29. Paaijmans, K. P., Read, A. F. & Thomas, M. B. Understanding the link between
malaria risk and climate. Proc. Natl Acad. Sci. USA 106, 13844–13849 (2009).
30. Hastings, D. A. & Dunbar, P. K. Global Land One-Kilometer Base Elevation (GLOBE).
Æhttp://www.ngdc.noaa.gov/mgg/topo/report/globedocumentationmanual.pdfæ
(National Geophysical Data Center, 1999).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements We thank T. L. Daniel, C. Martı
´nez del Rio, W. R. Rice and J.
Tewksbury for discussion, and S. L. Chown for sharing his data on latitudinal variation
in E. Researchwas funded in part by NSF IOB-041684 to R.B.H. andby an NSF Minority
Postdoctoral Fellowship to M.E.D.
Author Contributions M.E.D., G.W. and R.B.H. conceived the project, designed the
analyses and wrote the paper; M.E.D. and G.W. collated weather station data and did
temperature and metabolic rate calculations.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Readers are welcome to comment on the online version of this article at
www.nature.com/nature. Correspondence and requests for materials should be
addressed to M.E.D. (Michael.Dillon@uwyo.edu).
RESEARCH LETTER
706 | NATURE | VOL 467 | 7 OCTOBER 2010
Macmillan Publishers Limited. All rights reserved
©2010
METHODS
Weather station data. We downloaded all available ‘isd-lite’ weather-station data
(http://www.ncdc.noaa.gov/oa/climate/isd/index.php)
20
. From the initial 22,486
weather stations, we extracted a ‘restricted’ data set (3,186 stations) that sampled
throughout the entire IPCC standard reference period (1961–1990)
8
as well as up
to 31 December 2009, in all seasons, and at least every six hours. (We did not
include Antarctic stations.) This data set had a total of 493,256,415 temperature
measurements with an average of 8.8 temperature measurements per station per
day; it was used for Figs 1 and 2 and Supplementary Figs 1–4. For the ‘unrestricted’
data set, we included all stations (other than Antarctic) that had data for the above
time period, independentof the seasonality or frequency of sampling. This data set
had 5,561 stations (used for Supplementary Fig. 5).
Metabolic rate estimates. To estimate metabolic rates from temperature data, we
used empirically derived estimates of the coefficients (Eand b
0
) of the equation
relating metabolic rate to temperature and body mass for unicellular organisms,
multicellular invertebrates, amphibians and reptiles
12
. We excluded fish because
our data are air (not water) temperatures. We excluded birds and mammals
because their body temperatures will not match air temperatures. We excluded
plants because the temperature dependence of their metabolic rates differs fun-
damentally from that of animal ectotherms
31
.
Mean temperatures are expedient for analyses of the impacts of climate warm-
ing. However, because the relationship between temperature and metabolicrates is
inherently nonlinear, the use of mean rather than individual temperatures to
predict metabolic rates will induce spurious results
32
—an effect known as the
‘fallacy of the averages’
15
. To illustrate this fallacy, we recomputed metabolic rates
for geographical regions using mean annual temperatures (Supplementary Fig. 2)
for comparison with rates predicted from ‘instantaneous’ temperatures (Fig. 1b).
Note that the use of mean temperatures underestimates the predictedincreases in
metabolic rates
15
, and also de-emphasizes the impact of warming in the north
temperate zones relative to the tropics. Consequently, it is imperative to use high-
frequency temperature data and to compute metabolic rate separately for each
temperature measurement.
Our analysis includes temperatures that fall outside the normal tolerance range
of most organisms (that is, below ,0uC and above ,40 uC). We include these
values for analytical transparency, and because their inclusion is conservative for
our analyses. Eliminating negative temperatures (for example, substituting meta-
bolic rates at 0 uC for all temperatures below 0 uC) will have little effect because
metabolic rates are negligible at these extremely cold temperatures. Substituting
metabolic ratesat 40 uC for temperatures above 40 uCwill tend to reduce metabolic
rates at mid-latitudes where these hot temperatures occur in summer; this would
induce a downward bias in our predicted metabolic rates for the north temperate
zone. In other words, by not truncating metabolic rates to the normal tolerance
range (0–40 uC), we avoid a bias that would favourincreased metabolic rates in the
tropics.
Geographical coverage. Weather stations are not equally spaced across the world
(summarized in Supplementary Table 2, Supplementary Figs 1 and 5a). To adjust
for the uneven spatial distribution (and non-independence) of stations, we com-
puted mean temperatures and metabolic rates for all stations within 5ulatitude by
5ulongitude grid cells (see Fig. 2). We then calculated means and standard errors
for latitudinal regions (see Fig. 1) by weighting grid-cell means by interpolated
rectangular mid-cell areas
30
. We further tested the effects of weather station spatial
coverage on our conclusions by comparing analyses using restricted (3,186 sta-
tions) and unrestricted (5,561 stations) data sets. Our conclusions are robust and
independent of the data set used (Supplementary Fig. 5).
Assumptions. We assume that station temperatures match the temperature of a dry-
skinned ectotherm positioned in shade at 2-m height. Of course, mobile ectotherms
can often use behaviour (for example, microhabitat selection) to buffer body tem-
peratures against changes in air temperatures
23,24
. Thus our metabolic estimates
should be viewed as an estimate for a non-regulating, inert and exposed ectotherm.
We assume that a singlemetabolic curve (equation (1))applies to all ectotherms,
independent of latitude. However, some high-latitude ectotherms have relatively
raised metabolic rates, which may reflect metabolic compensation for temper-
ature
26
. To estimate whether metabolic compensation at high latitude might alter
latitudinal patterns (Fig. 1), we recalculated metabolic rates after substituting
extreme values of Eand b
0
. Specifically, we used very high values of E(0.76) and
b
0
(26.85) for the north temperate and Arctic areas, but very low E(0.50) and b
0
(15.68) for tropical areas (see Supplementary Table 1). Such extreme metabolic
compensation(these values span most of the known range of E)
26
does not alter our
conclusions (Supplementary Fig. 4).
Statistics. We used an ordered-heterogeneity test
21
to evaluate whether latitudinal
ordering of changes in temperature since 1980 predicts the latitudinal ordering of
changes in metabolic rate.
31. Reich, P. B., Tjoelker, M. G., Machado, J.-L. & Oleksyn, J. Universal scaling of
respiratorymetabolism, size and nitrogenin plants. Nature 439, 457–461 (2006).
32. Ruel, J. J. & Ayres, M. P. Jensen’s inequality predicts effects of environmental
variation. Trends Ecol. Evol. 14, 361–366 (1999).
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2010
